Insulin micropump

The primary limitation currently facing the development and use of these implantable micropumps is the precipitation of insulin on the pump components (7,11,14,31), although bicarbonate (33), certain serum components (25), or amino acides (32) may minimize or prevent insulin aggregation. These precipitates, if they accumulate, tend to occlude the flow channels in the pump and particularly the fine bore tubing or valves that are integral parts of these devices. Additional concern is related to the reliability and inherent safety of these devices and to their ability to use small, concentrated insulin reservoirs. 

The controlled-release micropump (Figure 2) is a recently invented device that uses the principles of membrane transport and controlled release of drugs to deliver insulin at variable rates (20,26). With a suitable supply of insulin connected to the pump, the concentration and/or pressure difference across the membrane results in diffusion or bulk transport through the membrane ). This process is the basal delivery and requires no external power source. Augmented delivery is achieved by repeated compression of the foam membrane by the coated mild-steel piston. The piston is the core of the solenoid, and compression is effected when current is applied to the solenoid coil. Interruption of the current causes the membrane to relax, drawing more drug into the membrane in preparation for the next compression cycle. 

To improve the delivery, 13-mm-diameter rate-controlling membranes held in a Swinnex filter chamber (Millipore Corp.) were inserted in the delivery line between the insulin reservoir and the micropump. The effective membrane area was 0.7 cm2. Membranes investigated were l- xm and 8- xm pore size polycarbonate filters (Nucle-pore Corp.), 0.45- xm cellulosic microporous filters (Amicon Corp.), Cuprophane PT-150 (from Ultra-Flow 145 Dialyser, Travenol Laboratories), and 0.2- xm and 1.2- xm pore size cellulose acetate filters (Schleicher and Schuell OE 66 and ST 69). 

Care had to be taken to fill the micropump with liquid since the presence of air bubbles in any of the lines would reduce the delivery rate. The portion of the pump below the membrane was filled with insulin-free phosphate buffered saline containing 0.5% (w/v) m-cresol (a preservative) via a tube connected to the pump outlet. When this portion was full, the membrane was laid onto the membrane support portion of the chamber, and the upper half of the chamber was reconnected to the controlled-release micropump. The upper half of the chamber constituted the 1-cm3 upstream reservoir for these experiments and was filled with radioactive feed solution through a needle inserted horizontally into the side of the membrane chamber. The top of the chamber was connected to a plastic three-way valve using the appropriate Luer-lok connections to permit filling. The valve was turned to seal the chamber and eliminate the pressure difference before the experiment. One of the ports of the valve was used... 

Insulin deposition in the controlled-release micropump is not expected to be important. While it was significant in one of the prototypes (Figure 4), changing the rate-controlling membrane from a hydrophobic polycarbonate filter to a hydrophilic Cuprophane or cellulose acetate membrane has apparently eliminated the problem. Although the situation may be different as longer-term experiments are performed, presumably the problems that may arise may relate more to the biological stability of the insulin reservoir than to insulin deposition. 

A number of concerns regarding open-loop delivery of insulin in general, and delivery of insulin by the controlled-release micropump, in particular, remain to be resolved. The optimum site of insulin delivery (intravenous... 

---